1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
// Copyright Materialize, Inc. and contributors. All rights reserved.
//
// Use of this software is governed by the Business Source License
// included in the LICENSE file.
//
// As of the Change Date specified in that file, in accordance with
// the Business Source License, use of this software will be governed
// by the Apache License, Version 2.0.

//! Fuses multiple `Join` operators into one `Join` operator.
//!
//! Multiway join planning relies on a broad view of the involved relations,
//! and chains of binary joins can make this challenging to reason about.
//! Collecting multiple joins together with their constraints improves
//! our ability to plan these joins, and reason about other operators' motion
//! around them.
//!
//! Also removes unit collections from joins, and joins with fewer than two inputs.
//!
//! Unit collections have no columns and a count of one, and a join with such
//! a collection act as the identity operator on collections. Once removed,
//! we may find joins with zero or one input, which can be further simplified.

use std::collections::BTreeMap;

use mz_expr::visit::Visit;
use mz_expr::{BinaryFunc, VariadicFunc};
use mz_expr::{MapFilterProject, MirRelationExpr, MirScalarExpr};

use crate::analysis::equivalences::EquivalenceClasses;
use crate::canonicalize_mfp::CanonicalizeMfp;
use crate::predicate_pushdown::PredicatePushdown;
use crate::{TransformCtx, TransformError};

/// Fuses multiple `Join` operators into one `Join` operator.
///
/// Removes unit collections from joins, and joins with fewer than two inputs.
/// Filters on top of nested joins are lifted so the nested joins can be fused.
#[derive(Debug)]
pub struct Join;

impl crate::Transform for Join {
    fn name(&self) -> &'static str {
        "JoinFusion"
    }

    #[mz_ore::instrument(
        target = "optimizer",
        level = "debug",
        fields(path.segment = "join_fusion")
    )]
    fn actually_perform_transform(
        &self,
        relation: &mut MirRelationExpr,
        _: &mut TransformCtx,
    ) -> Result<(), TransformError> {
        // We need to stick with post-order here because `action` only fuses a
        // Join with its direct children. This means that we can only fuse a
        // tree of Join nodes in a single pass if we work bottom-up.
        let mut transformed = false;
        relation.try_visit_mut_post(&mut |relation| {
            transformed |= Self::action(relation)?;
            Ok::<_, TransformError>(())
        })?;
        // If the action applied in the non-trivial case, run PredicatePushdown
        // and CanonicalizeMfp in order to re-construct an equi-Join which would
        // be de-constructed as a Filter + CrossJoin by the action application.
        //
        // TODO(database-issues#7728): This is a temporary solution which fixes the "Product
        // limits" issue observed in a failed Nightly run when the PR was first
        // tested (https://buildkite.com/materialize/nightly/builds/6670). We
        // should re-evaluate if we need this ad-hoc re-normalization step when
        // LiteralLifting is removed in favor of EquivalencePropagation.
        if transformed {
            PredicatePushdown::default().action(relation, &mut BTreeMap::new())?;
            CanonicalizeMfp.action(relation)?
        }
        mz_repr::explain::trace_plan(&*relation);
        Ok(())
    }
}

impl Join {
    /// Fuses multiple `Join` operators into one `Join` operator.
    ///
    /// Return Ok(true) iff the action manipulated the tree after detecting the
    /// most general pattern.
    pub fn action(relation: &mut MirRelationExpr) -> Result<bool, TransformError> {
        if let MirRelationExpr::Join {
            inputs,
            equivalences,
            ..
        } = relation
        {
            // Local non-fusion tidying.
            inputs.retain(|e| !e.is_constant_singleton());
            if inputs.len() == 0 {
                *relation = MirRelationExpr::constant(vec![vec![]], mz_repr::RelationType::empty())
                    .filter(unpack_equivalences(equivalences));
                return Ok(false);
            }
            if inputs.len() == 1 {
                *relation = inputs
                    .pop()
                    .unwrap()
                    .filter(unpack_equivalences(equivalences));
                return Ok(false);
            }

            // Bail early if no children are MFPs around a Join
            if inputs.iter().any(|mut expr| {
                let mut result = None;
                while result.is_none() {
                    match expr {
                        MirRelationExpr::Map { input, .. }
                        | MirRelationExpr::Filter { input, .. }
                        | MirRelationExpr::Project { input, .. } => {
                            expr = &**input;
                        }
                        MirRelationExpr::Join { .. } => {
                            result = Some(true);
                        }
                        _ => {
                            result = Some(false);
                        }
                    }
                }
                result.unwrap()
            }) {
                // Each input is either an MFP around a Join, or just an expression.
                let children = inputs
                    .iter()
                    .map(|expr| {
                        let (mfp, inner) = MapFilterProject::extract_from_expression(expr);
                        if let MirRelationExpr::Join {
                            inputs,
                            equivalences,
                            ..
                        } = inner
                        {
                            Ok((mfp, (inputs, equivalences)))
                        } else {
                            Err((mfp.projection.len(), expr))
                        }
                    })
                    .collect::<Vec<_>>();

                // Our plan is to append all subjoin inputs, and non-join expressions.
                // Each join will lift its MFP to act on the whole product (via arity).
                // The final join will also be wrapped with `equivalences` as predicates.

                let mut outer_arity = children
                    .iter()
                    .map(|child| match child {
                        Ok((mfp, _)) => mfp.input_arity,
                        Err((arity, _)) => *arity,
                    })
                    .sum();

                // We will accumulate the lifted transformations here.
                let mut outer_mfp = MapFilterProject::new(outer_arity);

                let mut arity_so_far = 0;

                let mut new_inputs = Vec::new();
                for child in children.into_iter() {
                    match child {
                        Ok((mut mfp, (inputs, equivalences))) => {
                            // Add the join inputs to the new join inputs.
                            new_inputs.extend(inputs.iter().cloned());

                            mfp.optimize();
                            let (mut map, mut filter, mut project) = mfp.as_map_filter_project();
                            filter.extend(unpack_equivalences(equivalences));
                            // We need to rewrite column references in map and filter.
                            // the applied map elements will be at the end, starting at `outer_arity`.
                            for expr in map.iter_mut() {
                                expr.visit_pre_mut(|e| {
                                    if let MirScalarExpr::Column(c) = e {
                                        if *c >= mfp.input_arity {
                                            *c -= mfp.input_arity;
                                            *c += outer_arity;
                                        } else {
                                            *c += arity_so_far;
                                        }
                                    }
                                });
                            }
                            for expr in filter.iter_mut() {
                                expr.visit_pre_mut(|e| {
                                    if let MirScalarExpr::Column(c) = e {
                                        if *c >= mfp.input_arity {
                                            *c -= mfp.input_arity;
                                            *c += outer_arity;
                                        } else {
                                            *c += arity_so_far;
                                        }
                                    }
                                });
                            }
                            for c in project.iter_mut() {
                                if *c >= mfp.input_arity {
                                    *c -= mfp.input_arity;
                                    *c += outer_arity;
                                } else {
                                    *c += arity_so_far;
                                }
                            }

                            outer_mfp = outer_mfp.map(map.clone());
                            outer_mfp = outer_mfp.filter(filter);
                            let projection = (0..arity_so_far)
                                .chain(project.clone())
                                .chain(arity_so_far + mfp.input_arity..outer_arity)
                                .collect::<Vec<_>>();
                            outer_mfp = outer_mfp.project(projection);

                            outer_arity += project.len();
                            outer_arity -= mfp.input_arity;
                            arity_so_far += project.len();
                        }
                        Err((arity, expr)) => {
                            new_inputs.push((*expr).clone());
                            arity_so_far += arity;
                        }
                    }
                }

                new_inputs.retain(|e| !e.is_constant_singleton());

                outer_mfp = outer_mfp.filter(unpack_equivalences(equivalences));
                outer_mfp.optimize();
                let (map, filter, project) = outer_mfp.as_map_filter_project();

                *relation = match new_inputs.len() {
                    0 => MirRelationExpr::constant(vec![vec![]], mz_repr::RelationType::empty()),
                    1 => new_inputs.pop().unwrap(),
                    _ => MirRelationExpr::join(new_inputs, Vec::new()),
                }
                .map(map)
                .filter(filter)
                .project(project);

                return Ok(true);
            }
        }

        Ok(false)
    }
}

/// Unpacks multiple equivalence classes into conjuncts that should all be true, essentially
/// turning join equivalences into a Filter.
///
/// Note that a join equivalence treats null equal to null, while an `=` in a Filter does not.
/// This function is mindful of this.
fn unpack_equivalences(equivalences: &Vec<Vec<MirScalarExpr>>) -> Vec<MirScalarExpr> {
    let mut result = Vec::new();
    for mut class in equivalences.iter().cloned() {
        // Let's put the simplest expression at the beginning of `class`, because all the
        // expressions will involve `class[0]`. For example, without sorting, we can get stuff like
        // `Filter (#0 = 5) AND (#0 = #1)`.
        // With sorting, this comes out as
        // `Filter (#0 = 5) AND (#1 = 5)`.
        // TODO: In the long term, we might want to call the entire `EquivalenceClasses::minimize`.
        class.sort_by(EquivalenceClasses::mir_scalar_expr_complexity);
        for expr in class[1..].iter() {
            result.push(MirScalarExpr::CallVariadic {
                func: VariadicFunc::Or,
                exprs: vec![
                    MirScalarExpr::CallBinary {
                        func: BinaryFunc::Eq,
                        expr1: Box::new(class[0].clone()),
                        expr2: Box::new(expr.clone()),
                    },
                    MirScalarExpr::CallVariadic {
                        func: VariadicFunc::And,
                        exprs: vec![class[0].clone().call_is_null(), expr.clone().call_is_null()],
                    },
                ],
            });
        }
    }
    result
}